pouët.net

Use of UE et cetera

category: general [glöplog]
because the line is so much blurrier.
added on the 2018-08-09 17:28:30 by msqrt msqrt
Quote:

i just want to be able to judge in the end if someone spent a lot time to work on a cool effect (code wise) or if someone "just" plugged an OBJ node into a geometry-to-particle-generator, twisted a few noise parameters and plugged that into a renderer.


Please don't want that, it's dumb. You're told what the compo platform is, and you're supposed to look at the prod and be able to tell if the result is impressive or not. You're supposed to know what can be done with today's tools and methods, and you're supposed to know what ideas have been done previously, and what is new. And you're supposed to know if you like the show artistically. If you can't tell what is impressive and what isn't on the modern Windows PC platform, then maybe you should learn more and watch more demos, or switch to an easier category?

Maybe I'm changing my mind about the "king of categories" status of big demos. It's actually the most demanding category for viewers.
added on the 2018-08-09 17:29:39 by yzi yzi
using a commercial engine is not a good criteria for a compo split. you can use unity and still implement your own material system with lots of custom shaders. on the other hand, you can link you "own engine" against filament and have a ready made pbr material system at hand. more commonly: pol still use audio engines like bass and don't write their mp3 players themselves. so far, noone complained? this debate is interesting, but ultimately just as pointless as the 3d accell thing back then. it comes down to looking beneath the surface if you really want to evaluate the effort and skill that went into something.
added on the 2018-08-09 17:34:14 by jco jco
Quote:
Why not split democompo.


Yeah! Separate the compos based on viewers! Separate compo for lamers who can't even tell what is good. Or maybe run the same demos twice, so you could win two prizes? I've always wanted to win the Lamers' Choice Award.
added on the 2018-08-09 17:35:07 by yzi yzi
Quote:
Why not split democompo.

Because barely anyone who enters those compos doesn't care, but to answer more seriously... hand-drawn vs. ray traced graphics produces (or at least back then produced) fundamentally distinct outcome and required very different skills which made the compo difficult to judge. Nowadays the two have converged so the distinction doesn't exist anymore. Tracked music vs. "streamed" is more difficult to say from an outside perspective, but I wager it was because people wanted it. But have you noticed what happened afterwards? People realized this was stupid. We have just one graphics compo and tracked music is relegated into novelty based on nostalgia. The graphical equivalent is the pixel graphics compo. So the implication on your end is that coding your own demos is a thing of the past and should be separated.
added on the 2018-08-09 17:39:23 by noby noby
I also guess that if there were a "demos using commercial engine" compo, that would invite lamers to release what in games is considered "asset flips", and nobody would give a shit about it. Consider that, even when using a commercial engine, it neither makes you a good demo coder, nor a good demo designer. Having a rather free competition for "everything that's executable on a decent PC" actually helps here.
Try it yourself, go animate a million particles with Unity in a way that makes sense.
I'll stick with my verdict that the debate is moot (while fun!).
added on the 2018-08-09 17:58:05 by jco jco
Also
a) lamers exist
b) lamers suck!
c) something that is fake can rule
d) not everything that rules is fake
e) there are things that do NOT rule because they're fake
f) ???
h) shader showdown!
added on the 2018-08-09 17:59:16 by jco jco
yzi: I am sorry,but I actually don't know all the features of Unity or UE to find out what's easy and what's not.

Let's see some examples:
1. DOF - up to this day many people show off with nicely implemented DOF with bokeh etc... I assume it's provided for you in UE without a hassle?
2. PBR - it's tricky to implement it by yourself, especially with IBL especially if your environment scene is changing. Now, if I see such effect in Unity demo, should I be impressed or not?
3. Vol.lights - let's assume it's there,but some coder adds extended version say under the water with some extra scattering effects - how would you even tell if he put a lot of effort or it was a 'standard package' if he doesn't mention this in the nfo file?
4. Subsurface scattering - many people praise Wander for cool SS tricks with milky fluid, what if some commercial engine has SS provided for you? Would you praise this demo so much? (BTW smoke scene is really awesome there!)

Then, let's say all those effects are there implemented for generic case and super easy to add to your scene (which is actually never the case, but nevermind). Now, if the demo doesn't contain any custom effects (and again, it's really hard to find out if it's not well exposed), what's left to judge (besides artistic part, which don't get me wrong, can be very demanding and can be consider as a crucial part of the demo - depending on the viewer) ?
added on the 2018-08-09 18:30:05 by tomkh tomkh
The upside is that if everyone just uses the readymade implementations, you'll be able to tell after a few prods :)

Quote:
I'll stick with my verdict that the debate is moot (while fun!).
Indeed -- much popcorn has been had.
added on the 2018-08-09 18:43:04 by msqrt msqrt
You compare what you see with all the other stuff you have seen that machine produce, in that size. Impressed? If yes, you vote for it.
added on the 2018-08-09 18:44:38 by yzi yzi
Haha yeah, it's that easy! As I said: try it yourself. Those aren't effects you can just switch on and voila, your scene looks great. You have to provide data for this stuff to work properly. Thing is, if you do not know what those features do, what requirements and restrictions they come with, you'll have a hard time making anything look good. Be it in a commercial or custom engine, or even when using a 3D modeling/rendering suite.
added on the 2018-08-09 19:00:20 by jco jco
the great gaping hole of a point this thread misses is that actually, in many cases, coding your own thing can be a massive advantage to both the creative process and the possible result. you can achieve something new and bespoke that would never be possible or would be a lot harder when living inside the confines of someone elses implementation in another tool.
you should code your own not because its "the demoscene way" or to live with some compo rule or audience expectation; you should code your own because you can actually achieve something better.

at least that *should* be true. i suspect the reason it isn't is the lack of creative abilities of many coders to make the most of their own implementations - they spend the time on writing the actual code but then use it in a shallow, uninteresting or ugly way.

given a coder who has the design skills too.. this is a different argument entirely.
added on the 2018-08-09 19:22:51 by smash smash
@smash, wise words. same goes for technical sound design on games too, as a side note ;). Coding is a game changer, it is control, it is a means for creative expression breaking the bounds of what is provided to you in form of "presets" or "templates".

On "newskool"-plaforms, we've been beyond the pure technical achievement for a while. It's no longer about "who can draw the most triangles to the screen" or the most dots or whatever. There, a paradigm shift, it happened, deal with it.

In more recent years I focused on animation, including technical animation, and direction. Aren't those valid skills you can present in demo form? (I already know the answer :P)

Personally, I appreciate it if someone investigates the intricacies of e.g. light interaction, regardless of this being done in a "commercial engine" framework or a weird 4k engine. The latter having the potential of being impressive because of clever compression schemes, the former being impressive nonetheless.

Wtf, just make cool demos ppl like to watch, doing realtime stuff with whatever you have available. I'll probably vote for a nice "amiga demo video player" over the boring "amiga not-a-video-player thing, because the latter left me super cold.
added on the 2018-08-09 19:42:15 by jco jco
(amiga demo video player including streaming of precalced triangle coordinates)
added on the 2018-08-09 19:42:47 by jco jco
Quote:
Let's see some examples:
1. DOF - up to this day many people show off with nicely implemented DOF with bokeh etc... I assume it's provided for you in UE without a hassle?
2. PBR - it's tricky to implement it by yourself, especially with IBL especially if your environment scene is changing. Now, if I see such effect in Unity demo, should I be impressed or not?
3. Vol.lights - let's assume it's there,but some coder adds extended version say under the water with some extra scattering effects - how would you even tell if he put a lot of effort or it was a 'standard package' if he doesn't mention this in the nfo file?
4. Subsurface scattering - many people praise Wander for cool SS tricks with milky fluid, what if some commercial engine has SS provided for you? Would you praise this demo so much? (BTW smoke scene is really awesome there!)

All these things are pure fun to make and really easy until you supposed to put them into other people's framework. I know studios where working on such tasks is considered privilege, and I mean it - eg only people succeeded in boring tasks are given an opportunity to work on shaders.
Also, all of these are absolutely known how to make.
Also, none of these will make your demo look great. If you don't trust me, go look at top 3 asm demos past several years, and imagine replacing these features if you will find them (which is unlikely, even for our ue4 demo =) with anything "less tricky" - you'll see that won't change much impression. Go figure. /shrug
added on the 2018-08-09 19:52:20 by ton ton
Stuff like "who has the best GI" or "who has the best physics simulations" are still legit things tho. Engines help to an extent but you can probably still do better by hand if you specialize for your cases and use a method that's just not feasible for games or other mainstream use cases (yet).
added on the 2018-08-09 19:54:48 by msqrt msqrt
Quote:
Stuff like "who has the best GI" or "who has the best physics simulations" are still legit things tho.

Yes, but it is worth mentioning that using gamedev engine for showing off cutting edge tech is a pure pain in the butt and really matches demoscene spirit (as it is one of the most painful ways one can imagine).
added on the 2018-08-09 20:02:58 by ton ton
(I mean, using other people's vanilla engine with monstrous number of abstractions and several abstraction layers which main feature is to cover all platforms you can imagine. And that's the engine we're talking about here)
added on the 2018-08-09 20:05:16 by ton ton
When scene tools like werkzeug where released many years ago, you'd see from time to time few releases from some sceners using these tools. But it didn't drastically change the scene, the majority still code demos from scratch or by programming their own tools. I am not afraid UE or Unity will be overused in the scene. Also, Poo-Brain demos made me more happy than the two UE demos I've experienced so far :)

I am kinda bored or too cynic to take part arguing on the issue, this is what I was doing when I was 20, I am old now. For one, I know I don't wish to use any ready made engines, since I want to explore coding things from scratch rather than designing a demo (I would need designers/artists/big motivated group anyway, and I suck at this). If others want to use whatever tools, I don't care.
added on the 2018-08-09 20:05:36 by Optimus Optimus
I am the best democoder in the world!!! Haha!!!

(j/k)
added on the 2018-08-09 21:24:31 by Adok Adok
Quote:
am kinda bored or too cynic to take part arguing on the issue, this is what I was doing when I was 20, I am old now. For one, I know I don't wish to use any ready made engines, since I want to explore coding things from scratch rather than designing a demo (I would need designers/artists/big motivated group anyway, and I suck at this). If others want to use whatever tools, I don't care.


High five. Though I'll keep releasing (winter this year again); I'd rather come up short than knowing I didn't try.
added on the 2018-08-09 22:26:39 by superplek superplek
Quote:
DOF, PBR, volumetric lights, subsurface scattering

I feel it should be pointed out that none of these should count as “effects” in 2018. They’re tools you can use to compose effects, but on their own they are nothing interesting. This seems to be core in the disconnect that’s happening here.

Side-note: boy, I can’t wait until the Meteoriks come around and this whole thing starts back up again.. :)
added on the 2018-08-09 23:41:15 by gloom gloom
please define "an effect". is it just the geometry? perhaps non-physical post-processing, glitches and stuff?

I'd definitely count those as effects, or at least interesting problems -- there are no definite realtime solutions to any of them, apart maybe PBR in a certain simple sense
added on the 2018-08-09 23:52:56 by msqrt msqrt
Quote:
Quote:
DOF, PBR, volumetric lights, subsurface scattering

I feel it should be pointed out that none of these should count as “effects” in 2018.


Dear Gloom, if this is what you think, you should really consider giving up on real-time computer graphics. Why do you torture yourself with such a difficult medium? Why not just use modern offline renderer? Take a look what you could do. It looks so much better than Wander or anything on a demoscene/UE/Unity that is barely implementing any of those things, which you refuse to call 'effects'.
added on the 2018-08-09 23:56:07 by tomkh tomkh
Also @ton

Quote:

Also, all of these are absolutely known how to make.


Thank you for clarification. I thought... or... maybe nevermind.
added on the 2018-08-10 00:07:29 by tomkh tomkh

login